of 8
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
1/17
Challenges with Cement Evaluation
What We Know & What We Don’t
SPE Webinar
July 11, 2012
Presented by David Stiles
ExxonMobil Development Co
Challenges with Cement Evaluation
If a tree falls in theforest and no one isthere to hear it, doesit still make a sound?
If an annulus is fullycemented but an acousticbond log can’t “hear” it, isthe cement really there?
Image courtesy of Schlumberger
Decisions, Decisions, Decisions
What do we do with log data?
• Select completion intervals
• Verify isolation of fresh water,
hydrocarbons & corrosive brines
• Decisions on remediation
• Continuous cementing
improvement
• Decisions on abandonments
• Satisfy regulatory requirements
What are the implications?
• Bypassed pay & lost reserves
• Lost production
• High WOR or GOR
• Stimulating out-of-zone
• Unnecessary holes in casing
• Difficulty cutting & pulling casing
• Unnecessary section milling
• Total loss of a well
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
2/17
What Do We Really Need to Know?
• Is the material behind pipe solid?
Isolation
No Isolation
• Is the material behind the pipe liquid?
SONIC BOND LOGS
Polling Question #1
What level of experience do you have interpretingbond logs?
A.What’s a bond log?
B.Bond logs come across my desk once in a while butthey’re just a bunch of squiggly lines to me
C.I have to look at a bond log once in a while but Idon’t have much confidence in my interpretation
D.I look at bond logs routinely and feel pretty goodabout my interpretations
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
3/17
Sonic Bond Log Physics 101
• SPE-1512, Oct. 1960 - A Sonic Method forAnalyzing the Quality of Cementation of Borehole
Casings – M. Grosmangin; P.P. Kokesh; P. Majani, Schlumberger
“The amplitude of a sound signal after it hastraveled in a firmly cemented pipe is only afraction of that recorded by the same device infree pipe.”
Conventional CBL
20 kHzTransmitter
3 ft.receiver
5 ft.receiver
Amplitude Curve
Variable Density Log (VDL)
Casing
Cement
Formation
How loud is the firstcasing arrival?
What does theentire waveformfrom all acousticpaths look like?
Polling Question #2
In what order do the acoustic signals arrive at the5 foot receiver?
A.Mud, Casing, Cement, Formation
B.Formation, Cement, Casing, Mud
C.Casing, Formation, Cement, Mud
D.They all arrive at the same time
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
4/17
Sonic Wave Propagation
Mud
Casing
Formation
Composite
Cement
Animation courtesy of Dr. Dan Russell, Grad. Prog. Acoustics, Penn State
VDL (Waveform) Processing
Raw Waveform
Half Wave Rectified and Shaded
Top View of Shaded Half Wave
Stacked Shading
Time (µs)
Typical CBL Presentation
1. Transit time
2. Gamma ray
3. CCL
4. Amplitude
5. VDL
1 2 3 4 5
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
5/17
CBL Assumptions
• The CBL sends out a monopole signal which measuresthe average circumferential amplitude
• The CBL assumes uniform cement behind the pipe
Animation courtesy of Dr. Dan Russell, Grad. Prog.Acoustics, Penn State
CBL Nomograph
• Tool specific chart
• Converts compressivestrength to amplitudeor attenuation andvice-versa
• Function of casingsize and thickness
Graphic Courtesy of Schlumberger
1.6 mV
3000 psi
10.1 db/ft
100 psi
3.5 db/ft
18 mV
Polling Question #3
What does >80% B.I. mean?
A.The acoustic impedance is >80% of what was anticipated
B.80% of the data is “Bad Information”
C.The cement is good
D.The cement is bad
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
6/17
Bond Index & Why is it Wrong?
• What is needed?
– Expected cement amplitude (E100%)
– Free pipe amplitude (Efree)
– Measured amplitude (EMeas )
• Bond index (BI):
log10(Emeas /Efree)
log10 (E100% /Efree)
• Conventional thought
– BI > 80% = “Good” cement
– BI < 80% = “Bad” cement
80% Channel = No Isolation
80% “Weaker” Cement = Isolation
Advanced Sonic Tools
• Advanced sonic tools have been developed that partially address
some of the limitations of the standard CBL
• Compensated Bond Tool (CBT)
– CBTs have two monopole transm itters and three monopole receivers
– They measure a compen sated signal attenuation independent of wellb ore fluid ortransmitter signal strength
• Sector Bond Logs (SBL)
– Standard monopole transm itter with omnidirectional 3 and 5 ft receiver s
– Plus 6 or 8 additional focu sed transmitter and receiver arrays for 60 o r 45 degree
azimuthal average amplitude
• Segmented Bond Tool (SBT)
– Actually an ultrasonic tool at ~100 k Hz but physics are similar to son ic logs
– Tool has six pads in direct contac t with the casing, each with a transmitter and
receiver that communicate with the adjacent pads
– This tool has 60 degree azim uthal resolution but otherwise works on similarphysics as the CBT
Typical SBT Presentation
1. Gamma ray
2. Amplitude of six sectors
3. Min & avg attenuation
4. Attenuation map
5. Standard VDL
1 2 3 4 5
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
7/17
Some Limits of Sonic Bond Logs
• Amplitude or attenuation readings only indicate what isdirectly acoustically coupled with the casing – it is not anindicator of isolation
• The VDL is only a qualitative indicator of the acousticcoupling at the cement/formation interface - it is not anindicator of isolation
• A microannulus or coated casing will cause acousticdecoupling
• Fast formations or cycle skip will cause misleadingamplitude readings
More Limits of Sonic Bond Logs
• Logging tools must be centered within specifications
• Evaluating low compressive strength cement isdifficult
• Evaluating gas cut or foamed cement is difficult
• Evaluating a thin cement sheath can cause problems
• Evaluation behind thick walled casing is difficult
• Interference in concentric casings can affect logging
• Channels cannot be identified with a standard CBLand even advanced sonic tools can’t identify small
channels
Q & A ON SONIC LOGS?
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
8/17
ULTRASONIC BOND LOGS
Polling Question #4
What parameters can an ultrasonic log (e.g. USIT,CAST-V) measure?
A.Azimuthal casing condition
B.Azimuthal acoustic impedance of material behind pipe
C.Azimuthal casing internal radius
D.Azimuthal casing thickness
E.All of the above
Ultrasonic Bond Log Physics 101
• Rotating pulse-echo transducer operating at 200 – 700 kHz
• Ultrasonic echoes produce four measurements
1. Pulse-echo amplitude⇒ casing condition
2. Pulse-echo transit time⇒ casing internal radius
3. Resonant frequency of casing⇒ casing thickness
4. Decay of the resonance⇒⇒⇒⇒ acoustic impedance of materialdirectly behind casing (MRayl = 106 kg.m-2.s-1)
Transducer Mud Casing Cement Formation
Wave animation courtesy of Dr. Dan Russell, Grad. Prog. Acoustics, Penn State
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
9/17
Ultrasonic Bond Log Physics 101 (cont.)
• Tools take 36, 72 or 100 readings per revolution
• Wellbore fluid properties are measured
– Fluid velocity (µs/in) – Fluid acoustic impedance in MRayl (106 kg.m-2.s-1 )
• Fluid density, velocity and impedance parameters arerequired inputs to measure the impedance of the materialbehind the casing
• Impedance readings are processed into a cement mapthat is color coded with the colors indicating the relativerange of the acoustic impedance readings
• Many logs use an arbitrary cut off point for cement of ≥ 2.6MRayl acoustic impedance
Acoustic Impedance of Various Materials
Tool accuracyis ± 0.5 MRayl
2.16 + 0.5 = 2.66
3.37 – 0.5 = 2.87
∆∆∆∆ = 0.21 MRayl
Typical Ultrasonic Presentations
1 2 3 4 5 6
1. Gamma ray
2. Casing Inspection
4. Min, max & avg
impedance
5. Percent bonded, liquid,gas and “micro-debonded”
6. Cement impedance map
3. Raw impedance data
1 8 9 6
7
7. Avg. impedance & tooleccentricity
8. Signal amplitude
9. VDL
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
10/17
Advanced Ultrasonic Tool
• The newest generation ultrasonic tool compliments theexisting pulse-echo (P-wave) measurements with flexural(S-wave) measurements
Animations courtesy of Dr. Dan Russell, Grad. Prog. Acoustics, Penn State
P-wave
S-wave
Advanced Ultrasonic Tool (cont.)
• The decay of the pulse-echo wave
provides acoustic impedance of the
material directly behind the casing
(same as a standard ultrasonic log)
• Flexural waves propagate along thecasing and generate shear and
pressure body waves in the cement
sheath
• Supported waves are dependent on
the material in the annulus
– Liquid supports only P-waves
– “Fast” cement only supports an S-wa ve
– “Slow” cement supports both S-waves &
P-waves
Far flexuralreceiver
Near flexuralreceiver
Flexuraltransducer
Pulse-echotransducer
Advanced Ultrasonic Tool (cont.)
• The combination of pulse-echo resonance and flexuralattenuation are used to discern between solid, liquid andgas directly behind the casing
Graphic Courtesy of Schlumberger
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
11/17
Typical Tracks on Advanced US Logs
1 2 3 4 5 6 71. Solid/Liquid/Gas Map
2. Flexural Attenuation
4. Signal Transit Time
5. Third Interface Echo Map
6. Velocity
3. Acoustic Impedance
7. Casing Eccentricity
Courtesy of Schlumberger
Always presented:
Presented only if a ThirdInterface Echo (TIE) isdetected:
Some Limits of Ultrasonic Bond Logs
• Maximum mud weight is ~16.0 lb/gal – less in NAF
• Logging tools must be centered within specifications
• A minor error in input logging fluid properties has a major
impact on measured impedance of the annular material
• Maximum casing thickness is ~0.75 inches
• Limits of tool resolution may make it hard to discern low
density, low strength cement from mud
• Poor casing condition will incorrectly affect the cement map
• A wet microannulus or coated casing will affect the
measurements, although less so than with sonic logs
• Depth of investigation is usually very small – no indication of
cement to formation interface
Q & A ON ULTRASONIC LOGS?
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
12/17
IMPROVING BOND LOGINTERPRETATION
Polling Question #5
What is the first thing you should do wheninterpreting a bond log?
A.Look for dark colors on the cement map
B.Check to see if the Bond Index is >80%
C.Review the well history
D.Call an expert for help
So How Do We Improve Confidence?
You can’t interpret a bond log in a vacuum
1.Determine what you expect the log to look like
– Know the objectives of the cement job
– Gather all pertinent data related to drilling and cementing of the
well
2.Perform thorough quality assurance checks
3.If the log is valid and it differs from what was expected orthe response can’t be explained…
– Use additional data processing or interpretation techniques
– Run additional logs as appropriate
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
13/17
Does the Log Look as Expected?
• What significant events occurred while drilling the section?
• What is the lithology?
• Was the job designed for good mud removal or was
channeling expected?
• Was the job executed as per design?
• Did significant events occur during the job (e.g. losses,
unplanned shut-downs, etc.)?
• Do actual job pressures match those predicted?
• What was the expected cement compressive strength at thetime of logging?
• Did significant events occur after cementing but prior to
logging?
Some Clues to Proper Interpretation
Borehole Quality
Pressure Matching
FormationArrivals
LithologyChange
Lithology Change
4.0
1.5
13.25
8.25
6.5
Compressive Strength
Courtesy of Baker
Pressure Pass to Confirm Microannulus
Zero Pressure Pass 1500 psi Pressure Pass
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
14/17
Log Quality Assurance – Is it Valid?
• Is the tool centered?
– Transit time on a CBL should be no more than 4 to 10 µsec/ft faster than thetime for a perfectly centered tool (only exception is f ast formations)
– For SBT, DTMN & DTMX curve separation should not exceed 4 to 5 µs /ft
• Does the repeat section match the main pass?
• If there is a free pipe section, is the amplitude within ± 10% of published
specs?
• Were casing size & weight input correctly?
• Were mud velocity & impedance measured & used for ultrasonic logs?
• Are any amplitude, attenuation or impedance curves less than zero?
• For logs with a cement map is the solid/liquid threshold correctly set for
expected cement and mud properties?
• For ultrasonic logs are there any processing error flags or bad
checkshot indicators?
Polling Question #6
Solids have a more consistent acoustic responsethan liquids
A.True
B.False
C.It depends
Different Presentation Improves Picture
• Liquids are relatively homogenous while solids are heterogeneous
Image Courtesy of the Portland Cement Assoc .Image Courtesy of Fir0002/Flagstaffotos
• The acoustic response to liquids is relatively flat while the responseto solids is wiggly – This is known as the Goodwin Presentation
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
15/17
Computer Post-Processing of the Data
If the following conditions are met:
1.All four standard deviations are higher
than set thresholds
2.The cement impedance is low (in thegas or liquid classes)
Then the current pixel is considered to
be of locally debonded set cement andis colored green
Transducer
“spot” size
Schlumberger USIT Micro-debonding
Graphic Courtesy of Schlumberger
Computer Post-Processing of the Data
• A statistical variationprocess to discern solidcrystalline structures, suchas cements, from fluids
• Analyzes the vertical andhorizontal rate ofimpedance change
Halliburton Advanced Cement Evaluation (ACE)
Courtesy of Halliburton
Summary
• A multitude of high impact decisions are based uponcement bond logs
• Sonic logs work on the principal that the amplitude of asonic signal as it travels through casing is relative to thedegree of cementation of the casing
• Ultrasonic logs work by producing a resonant frequency inthe casing, the decay of which can be used to determinethe acoustic impedance of the material behind the pipe
• All bond logging tools have many limits that preventcomplete certainty in their interpretation
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
16/17
Summary (cont.)
• Bond logs cannot be interpreted in a vacuum
• Log quality assurance checks are imperative
• Knowledge of the expected log response based onpertinent events associated with drilling and cementingthe well is required before interpreting a bond log
• Looking at data variance through a Goodwinpresentation or computer processing can decrease thedegree of uncertainty
FINAL Q & A
8/21/2019 120711 SPE Cement Evaluation Webinar Presentation
17/17
Nomenclature
DTMN: delta time minimum (for an SBT)
DTMX: delta time maximum (for an SBT)
GOR: gas-oil-ratio
MRayl: Mega Rayleigh
NAF: non-aqueous fluid
WOR: water-oil ratio